LIBRARY ı6ıul 0) - Cranfield University

8.4 Process monitoring and control
It is well established that through-the-arc sensing can be used for prediction,
monitoring and control of arc welding (see section 2.6). In this work, through-the-arc
sensing was used for assessing process stability and metal transfer, and for estimating
stand-off. The aspects involved in the stand-off estimation have been discussed in
section 8.3.2. The process stability and the mode of metal transfer were assessed
using the monitoring indices developed by Ogunbiyi [ref. 51]. These indices were
found to be useful in assessing the stability state of the process by flagging up
phenomena such as bad arc ignition, excessive spatter and excessive fuming and for
providing an objective way of characterising the mode of metal transfer (see section
4.2.1). They also give an indication of whether an inappropriate voltage for the set
wire feed speed is being used and whether an increase or decrease in the voltage level
is necessary to bring the process to an acceptable stability level (see Figure 4.2).
Therefore, they were used to develop the voltage controller (see section 4.2.1).
The monitoring of process stability was very important, since a reliable stand-
off estimation would depend on the process stability. The stand-off estimation models
(welding current cumulative differences and dip-resistance based models) were
implemented in such a way that they would only update the estimated values if a pre-
determined level of stability, estimated by using the Confidence of Bad Ignition model
(see section 4.2.3), was assured.
Two different controllers were integrated together in this work, one addresses
the positioning of the welding torch relative to the workpiece and the other addresses
the attainment and maintenance
of the process stability.
8.4.1 Position controller
The control of relative position (stand-off) between the welding torch and the
workpiece was performed using an independent workpiece positioning table. This
approach was adopted after extensive consideration of the options available for
implementing relative position control between the welding torch and the workpiece
during the robot program run-time (see sections 4.1.1 and 4.1.2).
The position control was performed by pre-weld searching the joint starting
point and in-process adjusting of the joint position so that it would follow the
programmed robot movement. The position controller was implemented making the
assumption that any programming errors had been corrected prior to welding and that
the robot and the welding cell had been properly calibrated. Although not critical, the
assumption that the robot had been calibrated was important to ensure that the joint
positioning errors would be due mainly to the component errors. It should be noted
that component errors are relatively small and normally restricted by the
manufacturing tolerances (see section 3.1.1.3). Also, the position controller is
constrained by the fact that the moving table has a limited range of movement,
implying that it can only compensate for errors within this range.
Only one degree of freedom was implemented in the position controller in
order to demonstrate that the positioning errors between the weld joint and the
programmed weld path can be minimised or corrected by adjusting the joint position.
194